Problems with csm_from_job_ini

[kirstybayliss]

I'm having some issues with hamlet running both locally and on the hazard workstation. On the workstation, I get the following error when trying to run hamlet:

I can also reproduce this error with the test at openquake/hme/utils/io/tests/test_io.py - on the workstation this fails in the same way. I guess it could be an issue with how hamlet is installed on the workstation. Running locally, this test fails for a different reason but at the same point:

I think this part at least is failing because the job.ini is not specified, which is now necessary (and we should update this test). Running locally with my own model and specifying a job.ini, I instead get the failure Kendra reported here #16 which Kendra's suggestion (now at line 598) fixes (but I cannot run hamlet locally for my model due to the very high number of rups in the composite source model).

I'm not sure if these are installation issues on the workstation side or if there's more going on here, but I wanted to write it all out in case anyone has any ideas and so we can figure out what actually needs changed to be able to consistently run hamlet.

[kirstybayliss]

Update: Antonio fixed the workstation issue, we needed a dbserver update (oq dbserver upgrade in case anyone else has this issue) :)
So the test still fails (because no job.ini) but hamlet runs for my model on the server.
I am still perplexed as to why I had issues locally given I have the same version of oq and hamlet